Daniel sent us this one — a three-part question that gets at something I think a lot of people are quietly wondering. When did clinicians first separate personality disorders from mood disorders, recognizing them as their own thing? How common are they really? And then the big one: narcissistic personality disorder has become this cultural fixation, but actual treatment remains elusive — so what breakthroughs, if any, are happening in the research? There's a tension here between how much the term gets thrown around and how little most of us understand the clinical reality, and I think that's where we need to start.
The gap between the pop-psychology "narcissist" and what clinicians mean by Narcissistic Personality Disorder is enormous. And that gap creates real problems — people self-diagnosing their ex-partners, armchair pathologizing, and meanwhile the people who actually have NPD are almost never the ones walking into a therapist's office saying "I think I'm a narcissist, please help." The core feature of the disorder works directly against treatment-seeking.
Which is the cruelest design flaw imaginable. The thing that defines the condition is the thing that prevents you from addressing it.
It's what clinicians call ego-syntonicity — the person experiences their traits as consistent with their self-image. They're not distressed by their grandiosity, their lack of empathy, their need for admiration. Other people are distressed. The patient is, by definition, mostly fine with it.
"The problem isn't me, it's everyone else's inability to recognize my exceptionalism." That's not a bug in the narcissistic mind, it's the operating system.
So to answer the question properly, we need to trace how clinicians even arrived at this category — when did they start seeing these as distinct conditions rather than just, say, someone with depression who happens to also be difficult? And that story is more recent than most people assume.
Let's dig into it. Where does this actually begin?
The pre-history is worth sketching quickly. You can find precursors going back to ancient humoral theory — the idea that personality traits reflected imbalances in bodily fluids. But the first clinical concept that genuinely anticipates personality disorders comes from Philippe Pinel in 1801. He described something he called "manie sans délire" — mania without delirium. Patients who had emotional disturbances and behavioral problems but no actual psychotic symptoms, no break from reality. They seemed rational but acted in ways that were deeply disordered.
"Mania without delirium." That's an elegant phrase for "you're not crazy, you're just impossible.
It was a important observation. Pinel was saying: these people aren't psychotic, they know what they're doing, and yet something is fundamentally wrong with how they function. That distinction — disordered personality without psychosis — was the conceptual seed. Then through the 19th century you get the concept of "moral insanity" from James Cowles Prichard and others, which was a mess of a term but captured the idea that the disorder was in character and behavior, not in cognition or mood.
Because of course the Victorians would frame a clinical condition as a moral failing.
And that framing haunts personality disorders to this day. The stigma is different from schizophrenia or depression. With personality disorders, there's always been this undercurrent of "is this really an illness, or are you just a bad person?" And that's part of why treatment has lagged so far behind.
When does this actually become formalized? When do we get the diagnostic category?
The inflection point is 1923. A German psychiatrist named Kurt Schneider — not to be confused with the later Schneider who worked on schizophrenia, this is a different Schneider — published a book called "Die psychopathischen Persönlichkeiten," "The Psychopathic Personalities." And this is the first systematic attempt to describe personality disorders as enduring patterns distinct from psychotic disorders and affective disorders. He wasn't using the term "personality disorder" yet — the language of the time was "psychopathic personality" — but he described types that map remarkably well onto what we'd later call antisocial, borderline, narcissistic, obsessive-compulsive personality disorders. His framework said: these are not episodic illnesses, they are stable patterns of behavior and inner experience that persist across the lifespan. That's the key distinction from mood disorders, which come and go.
Schneider draws the line in 1923. But it takes the DSM another fifty-seven years to actually separate them formally?
That's the frustrating part. The DSM-I in 1952 included something called "personality pattern disturbance," but it was lumped in with neuroses. The conceptual separation wasn't there. Clinicians still thought of these as essentially neurotic conditions — extreme versions of anxiety or depression, basically. DSM-II in 1968 didn't fix this. It took until 1980 — DSM-III — for the watershed moment.
What happened in 1980?
Robert Spitzer and the DSM-III task force created Axis II. This was a separate diagnostic axis specifically for personality disorders and intellectual disabilities, while Axis I covered everything else — mood disorders, anxiety disorders, psychotic disorders, substance use. For the first time, a patient could be diagnosed with both major depressive disorder on Axis I and borderline personality disorder on Axis II. The diagnostic manual was saying explicitly: these are different categories of illness that can co-occur but are not the same thing.
That was controversial at the time?
Because the comorbidity data made the separation seem almost artificial. This is one of the most important numbers in the whole history of this debate — a 2008 study by Zimmerman and colleagues in the Journal of Clinical Psychiatry found that up to 60 percent of patients with major depressive disorder also meet criteria for a personality disorder. When conditions overlap that much in the clinic, it's hard to tell whether you're looking at two separate things or two manifestations of the same thing.
The counterargument would be: if three out of five depressed patients meet criteria for a personality disorder, maybe you haven't actually separated them. Maybe you're just describing the same patient twice using different language.
That's exactly the critique that persisted for decades. And it's not a dumb critique — when the overlap is that high, you have to take it seriously. The counter-counterargument, which I think has largely won out, is that the longitudinal data tells a different story. Mood disorders are episodic — they come and go. Personality disorders are stable across decades. If you track patients over twenty years, the depression lifts and returns, but the personality pathology is constant. That's not what you'd expect if they were the same thing.
The temporal pattern is the giveaway. Depression is a visitor. Personality disorder is the architecture of the house.
That temporal distinction is what eventually drove the DSM-5 to introduce the Alternative Model for Personality Disorders in 2013. The categorical model — you either have NPD or you don't, you have borderline or you don't — was breaking down in practice. Too many patients fell in the "not otherwise specified" category. Too much overlap between supposedly distinct disorders. The dimensional model instead assesses two things: impairment in personality functioning, which covers identity and self-direction plus empathy and intimacy, and pathological personality traits across five domains — negative affectivity, detachment, antagonism, disinhibition, and psychoticism.
Instead of "do you have NPD, check yes or no," you get a profile: high on antagonism, low on empathy, impaired identity functioning. More like a personality spectrogram than a diagnostic label.
And this matters for treatment because it tells you what to target. If someone's primary problem is antagonism, that suggests different interventions than if the main issue is detachment. The categorical model said "you have NPD, good luck" — the dimensional model says "here are the specific domains of dysfunction, let's work on those.
Which brings us to the prevalence question. How many people are we actually talking about here?
The World Health Organization's World Mental Health Surveys, published in 2022, found a global lifetime prevalence of about 6.1 percent for any personality disorder. Cluster B disorders — that's antisocial, borderline, histrionic, and narcissistic — come in at roughly 2.8 percent globally. NPD specifically is estimated at about 0.5 to 1 percent of the general population. So it's rare, but not vanishingly rare. In a city of a million people, you're looking at five to ten thousand people who meet full diagnostic criteria.
Many more who have significant narcissistic traits without meeting the clinical threshold.
Right, and that's one of the misconceptions we need to address. The DSM-5 requires significant functional impairment across multiple life domains. Being self-centered, arrogant, or difficult is not NPD. The disorder requires that these traits cause clinically significant distress or impairment in social, occupational, or other important areas of functioning. The narcissistic CEO who's actually successful might not meet criteria. The narcissist whose relationships keep collapsing, who can't hold a job because they can't take direction, who burns through friendships — that's more like the clinical picture.
The cultural conversation has basically pathologized a personality style that, in many cases, is adaptive enough to not qualify as a disorder. The actual disorder is rarer and more devastating.
More treatment-resistant. That's the next part of the question, and it's where things get both discouraging and interesting. A 2020 meta-analysis in the Journal of Clinical Psychology found that the average dropout rate for NPD treatment is 42 percent. Nearly half of patients who start therapy don't finish it. And that's among the minority who seek treatment at all.
What drives the dropout? Is it the ego-syntonicity again?
When you don't believe there's anything wrong with you, therapy feels like an attack. The therapist challenges your grandiosity, and your response is "this person doesn't understand me" or "this person is threatened by me." But there's also a deeper dynamic. Otto Kernberg, who's been working on this since the 1970s, described what happens in the therapy room with narcissistic patients — they idealize the therapist initially, then inevitably feel disappointed or criticized, and then devalue and abandon the treatment. It's a predictable cycle.
Kernberg's 1975 book was "Borderline Conditions and Pathological Narcissism," right? That was the first major work to propose an actual psychodynamic treatment approach.
Before Kernberg, the therapeutic nihilism around NPD was almost total. The attitude was: these patients don't get better, they just burn through therapists. Kernberg said: no, there's a structure to this pathology, and if you understand the structure, you can work with it. His approach was to use the patient-therapist relationship itself as the tool — the patient's grandiosity, their need for admiration, their rage at perceived slights, all of that shows up in the room with the therapist, and that's where you work on it.
Instead of talking about the patient's failed marriage in the abstract, you're working with how the patient is treating you, right now, in the session.
That's the core of transference-focused psychotherapy, or TFP. And in 2021, we got the first major randomized controlled trial for TFP specifically for NPD. Diana Diamond and colleagues published it in the American Journal of Psychiatry. The effect size was Cohen's d of 0.78 compared to treatment as usual. That's a moderate to large effect — not a cure, but genuine, measurable improvement in narcissistic pathology. Patients showed reductions in grandiosity, improvements in interpersonal functioning, better emotional regulation.
78 is meaningful. For context, what's the typical effect size for antidepressant medication versus placebo?
For moderate to severe depression, antidepressant effect sizes tend to be in the 0.3 to 0.5 range depending on the study. 78 for a notoriously difficult-to-treat condition is impressive. Not a magic bullet, but real.
TFP is one breakthrough. What else has emerged?
The second major development is mentalization-based treatment, or MBT, adapted for NPD. This was originally developed by Peter Fonagy and Anthony Bateman for borderline personality disorder, but they've been adapting it for narcissistic pathology. A 2023 pilot study at the Anna Freud Centre in London found that MBT for NPD improved reflective functioning — that's the capacity to understand your own and others' mental states — and reduced interpersonal problems. Sixty-eight percent of patients showed clinically significant change at the twelve-month follow-up.
Walk me through what mentalization-based treatment actually looks like. How is it different from TFP?
TFP is psychodynamic and high-intensity — it works with the patient-therapist relationship in real time, surfacing and interpreting the transference. MBT is more structured and attachment-focused. The core idea is that people with personality disorders have impaired mentalization — they struggle to understand that other people have minds, intentions, feelings that are separate from their own. For NPD specifically, the impairment often takes the form of being unable to mentalize others except as sources of admiration or criticism. The therapist works on helping the patient pause and ask: "What might be going on in that person's mind right now? What might they be feeling that's different from what I'm feeling?" It sounds simple, but for someone with NPD, this is difficult work.
TFP says "look at what's happening between us right now." MBT says "let's practice understanding what's happening in other people's heads.
That's a reductive but fair summary. And then there's the third breakthrough — schema therapy. A 2024 systematic review in Clinical Psychology Review found schema therapy effective for NPD, particularly for what clinicians call the "vulnerable narcissist" subtype. Effect sizes ranged from 0.6 to 0.9 for reducing maladaptive schemas.
Define schema in this context. This isn't the same as a database schema.
Schemas in cognitive therapy are deep, enduring patterns of thinking and feeling about yourself and the world, usually formed in childhood. For NPD, the relevant schemas include things like emotional deprivation, defectiveness, entitlement, and insufficient self-control. The narcissistic presentation — the grandiosity, the need for admiration — is understood as a compensatory structure covering over deep feelings of shame and inadequacy. Schema therapy works by helping the patient access and heal those underlying schemas rather than just challenging the surface grandiosity.
That's a different theory of the case than TFP or MBT. You're saying the grandiosity is armor, and if you can heal what's underneath, the armor becomes unnecessary.
And for the vulnerable narcissist subtype — these are people who are grandiose in fantasy but fragile and hypersensitive to criticism in reality — this approach seems particularly well-suited. The overtly grandiose narcissist who's successful and charming might be harder to reach with schema therapy because they're less aware of the underlying pain.
We've got three evidence-based approaches now — TFP, MBT, schema therapy — all showing moderate to large effects in controlled studies. That's more than I expected. I thought we were still in "basically untreatable" territory.
The "basically untreatable" narrative is about twenty years out of date, but it persists in clinical training and in the culture. Part of the problem is that these specialized treatments aren't widely available. The Personality Disorders Institute at Weill Cornell trains clinicians in TFP. The Anna Freud Centre trains people in MBT. But your average community mental health center isn't offering any of this. So patients who could benefit can't access treatment, and the clinical impression that "nothing works" gets reinforced.
The availability problem creates a self-fulfilling prophecy. Treatment isn't available, so patients don't improve, so clinicians conclude treatment doesn't work, so nobody invests in making treatment available.
There's a pharmacological dimension worth mentioning. No FDA-approved medications exist for NPD specifically. But a 2025 open-label trial at the National Institute of Mental Health tested low-dose aripiprazole — 2.5 to 10 milligrams per day — and found modest reductions in grandiosity and aggression, about a 30 percent improvement on the Narcissistic Personality Inventory. That's not a breakthrough on the scale of the psychotherapy findings, but it's a signal that pharmacology might have a role as an adjunct.
Aripiprazole is an atypical antipsychotic, right? What's the hypothesized mechanism for NPD?
The hypothesis is that it modulates dopamine in a way that reduces reward-seeking behavior and impulsivity without the flattening effects of older antipsychotics. Grandiosity has a dopaminergic component — it feels good to feel superior. If you can modulate that reward system, you might reduce the drive toward grandiose self-enhancement. But I want to be clear — this was an open-label trial, no placebo control, small sample. It's promising but preliminary.
The pharmacological frontier is basically: we've got one small signal, no approved medications, and the real action is in psychotherapy.
Which makes sense given what personality disorders are. They're not chemical imbalances in the way depression might be. They're deeply ingrained patterns of relating to self and others, developed over decades. A pill isn't going to teach you how to mentalize. But a structured therapy might.
Let me ask you about something that's been tugging at me through this whole conversation. The STAR*D follow-up study from 2023 — that's the one finding that 47 percent of patients with treatment-resistant depression have unrecognized personality pathology. That number is startling. It suggests that a lot of people are being treated for depression when the real problem is something else entirely.
This is one of the most clinically important findings of the last decade, and it doesn't get nearly enough attention. STAR*D was the largest real-world study of depression treatment ever conducted in the United States. The 2023 follow-up analysis looked at patients who didn't respond to multiple antidepressant trials and found that nearly half had undiagnosed personality disorders. These were patients being treated for depression — sometimes for years — without anyone noticing that the depression was, in many cases, secondary to a personality disorder that nobody had assessed for.
You've got a patient with narcissistic personality disorder who keeps getting depressed because their grandiosity keeps crashing against reality. They lose jobs, relationships collapse, they can't sustain the admiration supply. Each crash looks like a depressive episode. They get antidepressants. The antidepressants don't work because the underlying problem isn't a serotonin deficit — it's a personality structure that generates repeated failures and losses.
That's exactly the pattern. And this is why the diagnostic separation matters so much clinically. If you treat the depression without addressing the personality disorder, you're treating the symptom, not the cause. The patient might feel temporarily better, but they're going to keep crashing. The STAR*D finding suggests that a significant portion of "treatment-resistant depression" is actually unrecognized personality pathology.
The actionable takeaway for clinicians is: if a patient isn't responding to depression treatment, screen for personality disorders. Don't just keep trying different antidepressants.
For patients and families, the takeaway is: treatment exists, but you need a specialized provider. These aren't conditions that respond to general supportive counseling. You need someone trained in TFP, MBT, or schema therapy. The Personality Disorders Institute at Weill Cornell and the Anna Freud Centre both offer directories of trained clinicians. It's not easy to find someone, but it's possible.
What about the general public? The people who've been marinating in the pop-psychology "narcissist" discourse for the last five years?
The main thing is distinguishing between narcissistic traits and Narcissistic Personality Disorder. Narcissistic traits are common and often adaptive. Confidence, self-regard, ambition — these exist on a spectrum, and moderate levels are healthy. NPD requires significant functional impairment. The DSM-5 says the impairment has to be in self-functioning — identity or self-direction — and interpersonal functioning — empathy or intimacy. Plus pathological personality traits in antagonism or grandiosity. It's not just being self-centered. It's a pervasive pattern that wrecks your life and the lives of people close to you.
The cultural conversation has basically taken a clinical category and turned it into a moral accusation. "He's a narcissist" means "he's bad, he's toxic, cut him out of your life." That's not a diagnostic process, it's a social sanction.
It creates real confusion. People come to therapy saying "I think my husband is a narcissist" when what they mean is "my husband is selfish and I'm unhappy." Those are valid concerns, but they're not the same as a clinical diagnosis. The clinical diagnosis requires a pattern that's stable across time and situations, that causes significant distress or impairment, and that can't be better explained by something else — substance use, another mental disorder, a medical condition.
Where is the field heading? The ICD-11 took effect in 2022 and eliminated categorical personality disorders entirely. No more "narcissistic personality disorder" as a distinct diagnosis under the ICD framework. Instead, you get a severity rating — mild, moderate, severe — plus trait domain qualifiers. Does this mean the term NPD is on its way out?
That's the open question. The ICD-11 model assesses the severity of personality dysfunction and then specifies the prominent trait domains — negative affectivity, detachment, dissociality, disinhibition, anankastia. So a patient who would have been diagnosed with NPD under DSM-5 might be diagnosed under ICD-11 with "severe personality disorder, prominent dissociality and negative affectivity." The clinical information is richer, but the familiar label disappears.
There's something lost and something gained there. The label "narcissistic personality disorder" has become so culturally loaded that maybe it's useful to retire it. On the other hand, the label helps patients and families find resources and communities.
The DSM has been more conservative. The DSM-5 kept the categorical disorders in Section II — the main diagnostic section — while placing the Alternative Model in Section III as an emerging model. The question is whether DSM-6, whenever it arrives, will follow ICD-11's lead and eliminate the categories or try to integrate the dimensional and categorical approaches.
I suspect the categories will hang on longer in the DSM because American clinical culture is attached to them. "Borderline personality disorder" is a whole identity and community now. "Narcissistic personality disorder" has entered the vernacular. You don't just retire those terms without a fight.
There are legitimate arguments for keeping them. The categorical labels facilitate communication — it's faster to say "NPD" than "severe personality disorder with prominent antagonism and impaired self-functioning." They also facilitate research, because you need consistent categories to build an evidence base. The dimensional model is more clinically accurate, but it's harder to study.
One thing I want to circle back to — you mentioned earlier that the old assumption was that personality disorders are lifelong and immutable. But the data doesn't actually support that, does it?
No, and this is one of the most important corrections to make. The Collaborative Longitudinal Personality Disorders Study — published in 2022 — found that 50 percent of patients with borderline personality disorder no longer met criteria after ten years. These conditions can remit. They're stable, but they're not immutable. And with targeted treatment, the trajectory improves further.
The "lifelong sentence" framing is wrong. And the more recent University of Zurich study from 2025 suggests that intervention can happen even earlier — that narcissistic traits in adolescence are modifiable with school-based mentalization programs.
That's the next frontier — early intervention. If you can reach adolescents who are showing narcissistic traits before those traits crystallize into a full personality disorder, you might prevent a lifetime of suffering. The Zurich study is small and preliminary, but the logic is compelling. Personality disorders typically emerge in adolescence and early adulthood. That's the window.
To pull this together: personality disorders were first systematically distinguished from mood disorders by Kurt Schneider in 1923, formally separated in the DSM-III in 1980, and are now being reconceptualized dimensionally in the ICD-11 and the DSM-5 Alternative Model. Global prevalence is about 6 percent for any personality disorder, with NPD specifically at about 0.5 to 1 percent. Treatment for NPD was considered nearly impossible for decades, but we now have three evidence-based psychotherapies — TFP, MBT, and schema therapy — with moderate to large effect sizes, plus preliminary pharmacological signals. The field has moved from "untreatable" to "treatable with difficulty," which is genuine progress.
The most important clinical insight might be the STAR*D finding — that nearly half of treatment-resistant depression may involve unrecognized personality pathology. If clinicians routinely screened for personality disorders in patients who aren't responding to depression treatment, a lot of people who are currently stuck might get unstuck.
Personality disorders are not character flaws or moral failings. They're deeply ingrained patterns that require sophisticated, evidence-based treatment. The field has moved from therapeutic nihilism to cautious optimism, and that's worth acknowledging even as we're honest about how hard the work remains.
Now: Hilbert's daily fun fact.
Hilbert: In the 1810s, Māori communities on New Zealand's South Island harvested a specific red seaweed known as karengo by scraping it from coastal rocks during low tide and sun-drying it into crisp sheets. Chemical analysis later revealed that this seaweed contains exceptionally high levels of taurine, a sulfonic acid more commonly associated with animal tissue, which gave the dried sheets a distinctively savory, almost meaty flavor profile unusual for a marine plant.
Seaweed that tastes like meat. Of course it does.
I have so many questions about the evolutionary purpose of taurine in seaweed, but I'm going to let them go.
If you found this episode valuable, please rate and review the show — it helps other people find us. Thanks to our producer Hilbert Flumingtop. This has been My Weird Prompts. Find us at myweirdprompts.
