Either all psychopathology is personality psychopathology or there is no such thing
What we call "personality disorders" may be better characterized as "interpersonal disorders."
While working with people with DSM/ICD personality disorder diagnoses, I often wonder: on what basis do we get to say that their personalities are disordered in a way that the personalities of people with mood, anxiety, and psychotic disorders are not? The most common personality deviations I see in the clinic are high neuroticism and low conscientiousness (disinhibition), but we almost never characterize these as personality disorders since they are so commonly seen in emotional and behavioral problems.
Recently, the question of what distinguishes the conditions we traditionally call “personality disorders” from other mental disorders has become the focus of intense discussion in the academic community. The stimulus for the current iteration of this debate has been the development of the Alternative Model for Personality Disorder (AMPD) in DSM-5 and its two main components, Criteria A and B, which are crucial for refining the conceptualization and diagnosis of personality disorders (PD).
Criterion A focuses on self and interpersonal dysfunction, and it was developed with significant influence from contemporary psychodynamic theory. Criterion B, on the other hand, is centered around quantitative differences in personality characteristics. It utilizes a dimensional trait model based on the Big Five personality traits (neuroticism, conscientiousness, agreeableness, extraversion, and openness to experience), which is well-studied and has strong empirical backing. Resultantly, it characterizes the personality disorder along five axes of negative affectivity, detachment, disinhibition, antagonism, and psychoticism.
Despite the intent for Criteria A and B to be complementary, the field has seen a growing body of literature that pits these criteria against each other, particularly in their ability to predict clinical outcomes and perform other psychometric tasks. In particular, an argument has emerged in certain circles that Criterion A is redundant at best and problematic at worst and should be eliminated in favor of diagnosing personality disorders on Criterion B.
The argument is presented most clearly in García et al. (2024), where authors propose discarding Criterion A (which focuses on maladaptive ways of viewing oneself and others) and relying solely on Criterion B (which identifies maladaptive personality traits):
“The evidence has failed to show that Criterion A is any different from pathological traits or that it adds incremental validity over them, and the theoretical assumptions that it represents the core features of PDs, differentiates PDs from other mental disorders, and is a better indicator of dysfunction lack support. Although the AMPD classification is psychometrically and theoretically superior to traditional categories, it would gain feasibility and coherence if Criterion A is discarded. Criterion B traits, together with an assessment of their real-life consequences, would be a less speculative measure of severity.” (García et al. 2024)
Hopwood (2024) and Zavlis and Fonagy (2024) have responded with persuasive and devastating demonstrations that this line of thinking leads to a stark either-or conclusion: either all mental disorders are personality disorders, or there is no such thing as a personality disorder. Both these responses are preceded by an earlier 2022 article by Aidan Wright and colleagues in which they argue that personality disorders are better conceptualized as interpersonal disorders.
(On a side note: These are all exceptionally well-written articles, and I highly recommend readers check them out. I am particularly fond of the Viewpoint format of article—I have written several myself and have commented on several on this blog in the past—and I have to say the piece by Christopher Hopwood is a prime example of the elegance and efficiency with which one can contribute to the literature with a viewpoint article.)
The proposal to define personality disorders in terms of their personality trait profiles seems unproblematic at face value, however
“it presumes that the association between personality traits and PDs is somehow special, if not unique, and that it is personality itself that is disordered as opposed to a circumscribed psychological dysfunction, delineated in scope or time. Three research literatures undermine these basic assumptions.” (Wright, et al. 2022; my italics)
First, research shows that personality traits are not uniquely associated with PDs; they also strongly correlate with other mental disorders, sometimes even more so.
“Large meta-analyses of big five personality trait associations with DSM PDs and clinical syndromes (e.g., Kotov et al., 2010; Samuel & Widiger, 2008) suggest there is nothing special about the relationship between personality traits and PDs relative to other disorders (Krueger et al., 2014). If anything, trait associations with clinical syndromes are stronger than with DSM PDs. For instance, Kotov and colleagues (2010) report that across anxiety, depressive, and substance use disorders, the mean effect size for neuroticism was r = .64, which is more than twice the average association between neuroticism and PDs (r = .31) reported in Samuel and Widiger’s meta-analysis. And similar patterns emerge for other traits (e.g., the average effect for conscientiousness across clinical syndromes [r = -.45] more than triples the mean association with PDs [r = -.13])” (Wright, et al. 2022)
Second, personality traits and other forms of psychopathology share the same structural space, suggesting that personality traits are not specific enough to define PDs uniquely.
“applying the quantitative empirical approach to adult psychopathology writ large results in a five-factor structure, including Internalizing, Detachment, Antagonism, Disinhibition, and Thought Disorder (Kotov et al., 2021; Wright & Simms, 2015), which are conceptually and empirically aligned with the basic Big Five model Neuroticism, Extraversion, Agreeableness, Conscientiousness, and Openness, respectively (Widiger et al., 2019). Thus, personality traits and all psychopathology, not just PDs, occupy the same structural space. At the same time, it must be emphasized that the two interpersonal factors of Antagonism and Detachment only emerge with the inclusion of PD diagnoses and/or features.” (Wright, et al. 2022)
Third, the assumption that PDs are more stable and pervasive than other clinical syndromes is not supported by scientific evidence, as both PDs and other mental disorders exhibit similar levels of stability.
These points suggest that personality traits alone cannot adequately describe the dysfunction that characterizes PDs. While the trait model of Criterion B is empirically robust, it fails to specifically define PD pathology, making it insufficient as a standalone model for PD diagnosis.
Hopwood articulates this quite well: a trait-based conceptualization of personality disorders amounts to the argument that there is no such thing as a personality disorder.
“Given that personality traits are not specifically associated with PD, that PD constructs cannot be distinguished from other kinds of psychopathology in terms of stability, and that trait-based models of PD do not specify the kinds of dysfunction that characterize PD, the argument that traits should be the primary criterion for diagnosing PD amounts to an argument that there is no such thing as PD. From this perspective, personality variance structures individual differences in the wide range of noncognitive mental health problems a person could have, inclusive of the kinds of problems that were listed in both the first and second Axes of DSM-III and DSM-IV, but there is no principled basis for distinguishing these Axes.” (Hopwood, 2024)
Zavlis & Fonagy put it this way: Either all mental disorders are personality disorders or there are no personality disorders.
“In other words, if personality disorder is merely a disorder of extreme personality traits, then all mental disorders can be cast as personality disorders because all mental disorders inherently entail personality extremeness. Of course, we all know better than to label every depressed, anxious, and autistic person as having a personality disorder simply because they score highly on measures of personality. Instead, we clinically understand that there is something different between these clinical cases and the classic “personality disorder” cases. And that something clearly is not reducible to personality traits.” (Zavlis & Fonagy, 2024)
Hopwood (2024) articulates this quite well: a trait-based conceptualization of personality disorders amounts to the argument that there is no such thing as a personality disorder. Zavlis & Fonagy (2024) put it this way: Either all mental disorders are personality disorders or there are no personality disorders.
Criterion A was developed to address the need to define and distinguish PDs by operationalizing interpersonal dysfunction. It draws from a broad range of clinical literature that identifies self-definition and interpersonal relatedness as central to personality development and functioning. The Levels of Personality Functioning Scale (LPFS) was created to measure these aspects, dividing them into specific sub-domains: Identity and Self-Direction for Self-Functioning, and Empathy and Intimacy for Interpersonal Functioning. These sub-domains are intended to link PD diagnosis with clinical practice by providing descriptions of the types of dysfunction seen in PD patients.
Criterion A has its own limitations. The LPFS, while useful, does not provide a comprehensive or coherent system of interpersonal functioning. It lacks specificity and fails to articulate the processes underlying interpersonal dysfunction. As a result, existing measures of Criterion A often reflect general psychiatric distress rather than specific PD pathology, making it difficult to distinguish PDs from other mental health issues.
There are two ways to tackle this problem with criterion A. The first is to appeal to theoretical account of personality functioning and development, and then to demonstrate that aberrations in personality development actually distinguish the conditions we call “personality disorders” from rest of psychopathology. The primary candidate for such a framework is psychodynamic developmental theory, but its scientific status remains subject to considerable controversy in the field of psychology and so far it has not been possible to empirically demonstrate a specific link between personality functioning and “personality disorders.”
The other approach is clinical and pragmatic: to focus on what is most clinically salient about this group of conditions. Just as we characterize bipolar disorder as a disorder of mood and schizophrenia as a disorder of thought, we can characterize conditions we currently call “personality disorders” in terms of interpersonal dysfunction—problems in relating to oneself and others—instead of hypothetical ideas about personality processes.
“The rationale for highlighting the interpersonal nature of PDs diagnostically is clear clinically. For instance, from a clinical perspective, what is important is that people with a PD diagnosis are perceived as “difficult” (Gibson & Ferrini, 2012; Kernberg, 2007; Moukaddem et al., 2017; Nakamura & Koo, 2017; Riddle et al., 2016; Stone, 2007; Treloar, 2009). They tend to have complicated attachment histories (Crawford et al., 2006) and they have difficulties maintaining stable social support systems (Beeney et al., 2018). Perhaps most importantly, people with PD diagnoses will have a difficult time developing an alliance in psychotherapy, a fundamentally interpersonal process (Bender, 2005), and difficulties persisting with treatment (Busmann et al., 2019). Clinicians generally expect patients with an anxiety or depressive disorder to establish an alliance and engage in treatment. In contrast, clinicians expect their interactions with patients who have PD diagnoses to be challenging. Thus, common treatments for PDs have added features designed to provide structure and establish boundaries, deal with ruptures and impasses, support clinicians, and set expectations for extended treatments with elevated probabilities of dropout or reversals (Bateman et al., 2015; Caligor et al., 2018; Oud et al., 2018). This connection between diagnosis and best therapeutic practice is the clinical reason to reformulate the psychopathology of PDs with reference to their interpersonal essence more specifically.” (Wright et al, 2022)
“we clinically understand that there is something different between these clinical cases and the classic "personality disorder" cases… that something appears to be a particular consequence of personality traits: maladaptive ways of relating. Clinically, the most striking feature of personality disorders has little to do with personality traits and much to do with the interpersonal consequences of those traits. These consequences include, but are not limited to, tendencies to view oneself and others in extreme and unstable ways (splitting), attachment and intimacy problems, and difficulties in forming therapeutic alliances or cooperating with others in any work-related capacity.” (Zavlis & Fonagy, 2024)
The key feature distinguishing “personality disorders” from other mental disorders is not merely the presence of maladaptive traits but the interpersonal consequences of these traits. Personality disorders are marked by difficulties in relationships, such as attachment issues, instability in self-perception, and problems forming therapeutic alliances.
If personality disorders are defined solely by extreme personality traits (Criterion B), then every mental disorder could be considered a personality disorder. Conversely, if the diagnosis is based on the negative interpersonal consequences of these traits (the intent of Criterion A), then personality disorders could be more specifically characterized as “interpersonal disorders.” This approach emphasizes the interpersonal difficulties often seen in PD patients. These interpersonal difficulties are central to the treatment of PDs and justify retaining the PD diagnosis to alert clinicians to these challenges.
Zavlis and Fonagy suggest that a map of personality traits and psychopathological domains would look something like this: high neuroticism linked to internalizing, emotional disorders; high disinhibition linked to externalizing, impulse regulation disorders; high psychoticism linked to thought disorders; and high antagonism and detachment linked to interpersonal disorders.
“… even when we categorise most psychopathologies in terms of personality, we still end up with the following (non-personality) categories: emotional disorders (high neuroticism), impulse disorders (high disinhibition), thought disorders (high psychoticism), and interpersonal disorders (high antagonism and detachment). Interestingly, the same themes emerge in factor-analytic studies of mental disorder domains, suggesting that the psychological themes of personality are inextricably linked to the psychological themes of psychopathology (Ringwald et al., 2023). These patterns suggest that there is nothing privileged about the association between personality and the traditional personality disorders. Instead, all mental disorders are invariably associated with personality themes that match their underlying pathologies. In that sense, all mental disorders both are and are not personality disorders.” (Zavlis & Fonagy, 2024)
All psychopathology both is and is not personality psychopathology.
Appendix:
In the Psychodynamic Diagnostic Manual-II, personality has a special place in the form of a separate “P Axis.” P Axis includes (1) level of personality organization and (2) personality style or type. Personality organization ranges from healthy, through neurotic and borderline, to psychotic levels. Personality types or styles represent clinically familiar personality styles or types that cross-cut levels of personality organization (such as depressive, obsessive-compulsive, narcissistic, and borderline personalities). PDM-II is explicit that these personality styles are prototypes and fuzzy sets that can be approximated to varying degrees, and not distinct categories. The concept of personality style does not inherently connote either health or pathology, but rather core psychological themes and organizing principles. In the PDM-II, there is no hard and fast distinction between a personality type or style and a personality disorder:
“The term “disorder” is a linguistic convenience for clinicians, denoting a degree of extremity or rigidity that causes significant dysfunction, suffering, or impairment. One can have, for example, a narcissistic personality style without having narcissistic personality disorder.” (PDM-II, p 27)
In other words, according to the psychodynamic perspective, everyone has a particular personality organization and particular profile of personality characteristics, and this shapes the psychopathology they experience and exhibit. Personality is relevant to and colors all psychopathology.
See also:
Excellent post! It introduced me to the whole debate. I am saving it for further study and will assign my resident to read it.
(About Las Meninas… Of all things I concluded my last Substack post with commentary about this unbelievable painting, and a reproduction of it as well! ) Thanks
In nowhere here does anyone use the word “Ego-syntonic”, what is otherwise standardly used to differentiate personality disorders (PD) from non-PDs. This heuristic has remained unchanged for over a century.
Step 1. Does the individual in question have a recurring series of “pathological traits”, or more bluntly, “traits” that can be characterized as “negative” according to the researcher in the ivory tower?
(We can replace the word “processes” or “states” in place of “traits” so long as they’re measurable).
Step 2: Do these same “traits” highly reliably precede all of their distress or dysfunction? Is there a robust causal link between these so-called pathological traits and so-called pathological events or dynamics involving other individuals, regardless of context?
If yes, this unambiguously points us toward the self. Because such outcomes are a-contextual across space and time (e.g., other people), and only selves can persist across time (so claim most enactivists), then the so-called “problem” must be the self itself. And unless skeptics are seriously prepared to defend metaphysical idealism or a “Universal Self” instead, we mean the self that organizes that particular body: Bob. That is, what others have called the personality, the soul, the psyche, and so on, also means the self. And what the psychoanalysts in the Appendix of Aftab’s post here have called “organization”. I will just collapse all these words as form, since form and organization are literally the same word anyways (see: Aristotle). Even if one demands a neurological definition of a self, such as Dan Zahavi’s “minimal self” or Georg Northoff’s “model of the self”, both emphasizing the “higher level” default and salience networks, those too are emergent forms of matter.
Step 3: Are these same “pathological traits” ego-syntonic to the individual?
Step 4: if such “traits” (or processes or states or dynamics) are ego-syntonic to the extent we can charitably say the individual has intentionally used them to actualize their (“problematic”) values across time, then the necessary criteria for personality disorder is met. If no, then no personality disorder may not be diagnosed at all. Not even in future taxonomies or models of personality disorders, unless they also provide a comprehensive and eternal theory of the psyche, self, soul, personality, ego, and so on that’s also compatible with systems biology (Aristotle’s hylomorphism already does this). The Big 5 does not qualify as it is not a comprehensive mind-body theory in the first place. It's explicitly "a-theoretical" instead.
I agree with Zavlis & Fonagy’s suspicion of defining personality disorders as “lists of traits”.
At best, traits are properties of emerging processes (rather than stand-alone substances or objects). For example, “Bob’s neuroticism” is not a distinct object with mass, charge, or spin that occupies space and time, atleast not “independent of Bob”. No one has ever “found where neuroticism is located” outside of a dataset because it is emergent in causal terms (Zachar & Krueger, 2013, p. 904). And so-called neural correlates “of neuroticism” are not actually “of neuroticism” in the first place, they are correlates of particular brains that are organized in particular ways in particular bodies. Aka, Bob's current form. Re-naming neurons to some “thing” other than neurons, nonetheless independent of it’s organ-ization (aka: form), is a category error. Likewise, shoe-horning a population-specific emergent property, such as neuroticism “inside of the brain” is also a fallacy, so complains Denny Borsboom (again, in Zachar & Krueger, 2013, p. 905).
Since this very logic came out of the book called “The Oxford Handbook of Philosophy and Psychiatry”, then the same can be said of “borderline personality disorder”. PD researchers might “find” neural correlates of particular brains whose’ organizing host has been labeled with BPD. However, such researchers are not actually describing BPD. They are describing an ensouled body to Aristotileans, a “personality” to a PD researcher, a “psyche” to “Psyche-ologists” and “Psyche-iatrists”, or an “organization” to a “psyche-analyst”, and so on. Neither psychiatrists, psychologists, counselors, or researchers have conversations with mental disorders. They instead have conversations with psyches. Indeed, it is because souls, psyches, egos, selves, personalities, and hence forms (all functionally the same word) are substantial given it is they (and nothing else) that organ-izes matter into orga-nelles and org-ans and organ-systems, and so on, and thereby enables and constrains (top-down), the strong emergence of mental processes.
Perhaps if they focused on the correct Greek and Latin roots of the words, rather than decade-specific statistical fads or “purely data driven” nonsense, we might avoid these issues. As it turns out, when you throw away all theory and basic assumptions, just as these PD researchers have done by appealing to the staunchly a-theoretical Big 5 paradigm, you then throw away interpretation and millennia's of insights that we killed kings over.
Zachar, P., & Krueger, R.F. (2013). Personality disorder and validity: a history of controversy (p. 889-910). In (Eds) Fulford, K.W.M., Davies, M, Graham, G., Sadler, J.Z., Stanghellini, G., & T., Thorton. (2013). The Oxford handbook of philosophy and psychiatry. Oxford: Oxford University Press.