In recent years, dimensional models of personality disorder have become increasingly influential, particularly among trait psychologists and personality researchers. Two prominent examples are the ICD-11 personality disorder framework and the Alternative Model for Personality Disorders (AMPD) included in the appendix of DSM-5.
Although these models differ in important ways, they share the assumption that personality pathology is best understood dimensionally rather than as a set of distinct disorders. According to this view, conditions such as borderline, narcissistic, and antisocial personality disorder reflect different configurations of maladaptive traits and severity rather than fundamentally separate syndromes.
Although the AMPD retains the traditional personality disorder categories alongside dimensional trait ratings, this hybrid structure was largely a compromise designed to ease the transition away from the categorical model (see Gunderson, 2013). Many proponents of dimensional approaches ultimately favor a fully dimensional system in which the existing categories are substantially deemphasized or abandoned altogether.
I wish here to raise three objections to dimensional models of personality disorder.
1. Dimensional models do not account for differences in diagnostic validity between personality disorders
One of the enduring problems with the DSM system is that it implicitly treats all diagnoses as equally valid or “real.” Thus, schizophrenia, bipolar disorder, and obsessive compulsive disorder are placed alongside far more weakly validated conditions as if all represent the same degree of empirical support (Ghaemi & Ruffalo, 2026).
When it comes to personality disorders, two diagnoses largely satisfy classic Robins and Guze (1970) criteria for diagnostic validity in psychiatry: borderline and antisocial (Gunderson, 2013; Paris, 2020; Ruffalo, 2026). Most of the others have weaker empirical validation. Yet, dimensional models of personality pathology tend to flatten these important distinctions in phenomenology, longitudinal course, and treatment response.
This fact ultimately led Gunderson (2013) to suggest listing borderline and antisocial disorders separately from the rest of the personality disorders, which are much more readily dimensionalized. In the end, this may be the most viable solution to preserve these two well-validated and clinically important diagnostic categories.
2. Dimensional models implicitly treat all personality disorders as if they represent the same kind of thing
In their effort to dimensionalize personality function, trait-based approaches essentially collapse fundamentally different forms of psychopathology into variations of a single underlying construct called “personality pathology.” From a classic psychiatric perspective (Guze, 1992), this is a major ontological assumption rather than an established fact.
For instance, schizotypal, paranoid, and some schizoid presentations have historically been understood as lying on the schizophrenia spectrum rather than belonging to the same class of disorders as borderline and antisocial personalities. In the late 1970s, a great deal of effort was expended to distinguish borderline personality from schizotypal personality and so-called “borderline schizophrenia” (Spitzer et al., 1979), reflecting the recognition that these conditions represented importantly different forms of psychopathology.
Dimensional models risk obscuring these important distinctions by treating all personality disorder categories as expressions of a single personality pathology dimension when they may in fact represent importantly different forms of psychopathology.
3. Dimensional models overstate the significance of fuzzy diagnostic boundaries and mistake them for evidence against the reality of distinct syndromes
Personality Disorders Essential Reads
Advocates of dimensional systems often point to the absence of perfectly discrete boundaries between disorders as evidence against categorical diagnosis. Yet the same problem exists throughout general medicine. As Roth and Kroll (1986) observed, “the same vague, uncertain area between health and disease is present in relation to the commonest types of physical illness.” Diseases such as hypertension, diabetes, and epilepsy are often fuzzy around the edges, but this does not negate the reality of the categories themselves.
The same principle applies to personality pathology. Borderline, antisocial, and narcissistic conditions may overlap in certain traits or symptom domains, but overlap alone does not eliminate the existence of meaningful diagnostic distinctions. Disorders can be fuzzy at the boundaries while remaining coherent at the center. Moreover, symptoms in personality pathology do not cluster randomly. They reliably coalesce into recognizable syndromes with characteristic patterns of affect, attachment, aggression, identity disturbance, and interpersonal functioning (Gunderson, 1984).
The existence of ambiguous or intermediate cases therefore does not undermine the reality of distinct personality syndromes any more than it undermines the reality of disease categories elsewhere in medicine.
Conclusion
Dimensional models of personality disorder may capture certain broad features of personality pathology, particularly severity and maladaptive trait variation. However, they often do so at the cost of obscuring important differences in diagnostic validity, phenomenology, and relationship to other forms of psychopathology.
Psychiatry has long depended on the recognition of distinct clinical syndromes that emerge through characteristic patterns of symptoms, course, familial aggregation, and treatment response (Guze, 1992). Borderline and antisocial personality disorders, in particular, appear to possess this kind of diagnostic coherence.
The existence of overlap, ambiguity, or dimensional variation does not negate these realities any more than it does elsewhere in medicine. A more viable approach may therefore be one that incorporates dimensional insights without abandoning the reality of meaningful diagnostic categories.