APA Citation
Kernberg, O. (1992). Aggression in Personality Disorders and Perversions. Yale University Press.
Summary
In this comprehensive work, Otto Kernberg examines how aggression—both overt and covert—shapes personality pathology. Drawing on decades of clinical experience treating severe personality disorders, Kernberg argues that unintegrated aggression is central to understanding narcissistic, borderline, and antisocial presentations. The book explores how early developmental failures lead to splitting, where aggression cannot be integrated with love, creating the characteristic instability and destructiveness of these conditions. Kernberg particularly illuminates the "borderline-narcissistic" combination, where patients alternate between desperate vulnerability and contemptuous grandiosity depending on whether their attachment system feels threatened. This analysis provides clinicians with crucial frameworks for understanding patients who seem to change personalities based on context.
Why This Matters for Survivors
For survivors of narcissistic abuse, Kernberg's work explains the bewildering experience of seeing your abuser shift between vulnerability and cruelty. Understanding that these aren't two different people but one person with unintegrated aggression helps make sense of the chaos—and confirms that the cruelty was real, not imagined.
What This Research Found
Aggression as the organising principle. Otto Kernberg’s seminal work argues that understanding personality disorders requires understanding aggression—not just overt violence, but the full spectrum from rage to contempt to subtle exploitation. In narcissistic, borderline, and antisocial presentations, aggression has not been properly integrated with loving feelings during development. This creates the characteristic instability: people become ‘all good’ or ‘all bad’ because the patient cannot hold mixed feelings.
The mechanism of splitting. Splitting is the psychological defence where contradictory feelings are kept apart rather than integrated. Kernberg explains that healthy development allows us to recognise that the same person can frustrate us and love us, disappoint us and support us. When early development is severely disrupted, this integration fails. The narcissistic personality disorder sufferer splits their self-image (grandiose vs. worthless) and their perception of others (idealised vs. devalued). Aggression lives in the ‘bad’ compartment and emerges when that compartment is activated.
The borderline-narcissistic combination. Kernberg identifies a common clinical presentation that confuses many therapists: patients who alternate between borderline vulnerability and narcissistic grandiosity. When attachment feels threatened, borderline features emerge—desperate clinging, self-harm, identity confusion, terror of abandonment. When stabilised, narcissistic defences appear—devaluation, contempt, entitlement, exploitation. The person seems to become someone entirely different, leaving partners and therapists feeling ‘whipsawed.’ This isn’t acting; it’s a genuine state-dependent shift in psychological organisation.
Implications for severity. Kernberg places personality disorders on a spectrum of severity based on identity integration, defence mechanisms, and reality testing. Malignant narcissism—combining narcissistic grandiosity with antisocial exploitation, paranoid features, and ego-syntonic aggression—represents the most severe end. These individuals may enjoy cruelty and feel justified in harming others. Understanding this spectrum helps clinicians assess dangerousness and set realistic treatment expectations.
Why This Matters for Survivors
Your confusion makes sense. If you’ve been in a relationship with someone who seemed to transform from a loving partner into a cruel stranger and back again, Kernberg’s framework explains what you experienced. You weren’t imagining things or being ‘too sensitive.’ The person genuinely operates from different psychological states—vulnerable and seeking comfort in one moment, aggressive and contemptuous in another. Both are real; the problem is they cannot be integrated into a coherent whole.
The cruelty wasn’t about you. When a narcissistic person devalues and discards you, it feels devastatingly personal. Kernberg’s analysis reveals it’s actually impersonal—you activated their ‘bad object’ compartment, the repository of all the aggression and contempt they cannot integrate with their idealised self-image. Someone else in your position would have received the same treatment. Understanding this doesn’t make the pain less real, but it can help you stop searching for what you did wrong.
You were dealing with something structural. Many survivors spend years trying to be ‘good enough’ to prevent the shift from idealisation to devaluation. Kernberg’s work clarifies that this shift is structurally inevitable—it’s not caused by anything you did or failed to do. The narcissistic defenses cannot sustain idealisation indefinitely. When disappointment occurs (and it always does, because no one is perfect), the shift happens. You could not have prevented it through better behaviour.
Recognising covert aggression. Kernberg illuminates how aggression in covert narcissism can be subtle—chronic disappointment, emotional unavailability, passive undermining, criticism disguised as concern. If you struggled to explain what was so harmful about a relationship that looked acceptable from outside, this framework validates your experience. Aggression that doesn’t involve raised voices or thrown objects is still aggression.
Clinical Implications
State-dependent intervention. Kernberg emphasises that treatment must be calibrated to the patient’s current psychological state. Warm empathic validation appropriate for a patient in borderline distress may reinforce grandiosity if delivered when narcissistic defences are active. Confrontation appropriate for narcissistic devaluation may be destabilising during borderline vulnerability. Clinicians must continuously assess which state is operating and adjust accordingly.
The centrality of the therapeutic relationship. Kernberg’s Transference-Focused Psychotherapy (TFP) uses the therapeutic relationship as the primary arena for change. Patients inevitably recreate their splitting patterns with the therapist—idealising, then devaluing; demanding, then withdrawing. Rather than correcting these patterns through instruction, TFP helps patients observe them in real-time within a relationship that survives them. This provides the experience of integration that was missing in development.
Assessing the severity spectrum. Not all personality disorders are equally treatable. Kernberg’s framework helps clinicians distinguish between patients who can engage in intensive psychotherapy and those whose aggression, antisocial features, or paranoia make this dangerous or impossible. Realistic treatment planning requires honest assessment of where a patient falls on the severity spectrum—and whether the clinician has the training and resources to work with that level of pathology.
Countertransference as data. Clinicians working with personality disorders often feel confused, manipulated, devalued, or idealised themselves. Kernberg treats these reactions not as problems to suppress but as valuable information about the patient’s internal world. If the therapist feels whipsawed between being the world’s best clinician and an incompetent failure, this mirrors the patient’s experience of relationships. Understanding countertransference helps clinicians maintain effectiveness rather than being destabilised.
Long-term treatment requirements. Kernberg is clear that meaningful change in severe personality disorders requires years of specialised treatment. The structural deficits that create splitting developed over decades and cannot be resolved quickly. While symptom reduction may occur faster, genuine personality reorganisation—the capacity to integrate aggression with love, to experience others as whole people—requires sustained work. Clinicians should set realistic expectations and avoid promising quick fixes.
Broader Implications
Understanding Domestic Violence
Kernberg’s framework illuminates the cycle of abuse in intimate relationships. The alternation between honeymoon phases and abuse mirrors the split between idealisation and devaluation. Abusers may genuinely believe their partner is wonderful during idealisation and genuinely experience them as deserving contempt during devaluation. This doesn’t excuse the behaviour, but it explains why traditional anger management often fails—the aggression isn’t a skill deficit but a structural problem in personality organisation.
Workplace Dynamics
The borderline-narcissistic pattern Kernberg describes frequently appears in organisational settings. Leaders may idealise new employees, showering them with praise and opportunity, then dramatically devalue them when inevitable disappointments occur. Understanding this pattern helps HR professionals and colleagues recognise that the problem isn’t the employee’s performance but the leader’s psychological structure. It also explains why narcissistic leaders often have high turnover and why some organisations develop systematically destructive cultures.
Custody and Family Court
Family courts frequently encounter parents with personality disorders. Kernberg’s framework helps evaluators understand why one parent might make extreme accusations against the other—the splitting that creates ‘all good’ self and ‘all bad’ other feels genuinely real to the person experiencing it. It also illuminates why some parents struggle to maintain consistent co-parenting: the same person can be idealised when cooperating and demonised when setting boundaries. Courts benefit from personality disorder literacy.
Cultural and Political Manifestations
The splitting Kernberg describes operates not just individually but collectively. Political movements that divide the world into pure allies and evil enemies, that cannot tolerate nuance or complexity, may reflect large-scale splitting dynamics. Understanding this helps explain the intensity of political polarisation and suggests that simply providing information won’t resolve it—the division serves psychological functions that must be addressed.
Therapeutic Training and Supervision
Kernberg’s work highlights that treating severe personality disorders requires specialised training most clinicians don’t receive. The standard therapy training may leave practitioners overwhelmed or harmed when working with patients who split, devalue, and project aggression. This has implications for training programs, supervision requirements, and the ethical responsibility to refer patients requiring expertise one doesn’t possess.
The Limits of Insight-Oriented Treatment
While Kernberg developed a psychodynamic treatment, his framework also illuminates when insight-oriented approaches may be inappropriate. Patients with significant antisocial features may use therapeutic insight to more effectively manipulate rather than to change. Patients with severe paranoia may experience interpretation as attack. Clinical humility requires recognising that some presentations require containment, limit-setting, or even no treatment rather than exploratory therapy.
Limitations and Considerations
Psychoanalytic concepts resist operationalisation. Terms like ‘splitting,’ ‘projective identification,’ and ‘primitive idealization’ are difficult to measure objectively. While clinically useful, they can be applied too liberally, potentially pathologising normal variation. Kernberg’s framework is best understood as a clinical lens rather than a scientific theory in the strict sense.
Treatment accessibility is limited. Transference-Focused Psychotherapy requires extensive training most therapists don’t have, and treatment lasting years that most patients cannot afford or access. The gap between what works and what’s available is substantial. Kernberg’s insights may be most useful for assessment and understanding even when full TFP isn’t possible.
The population studied is specific. Kernberg’s framework was developed through work with patients in intensive treatment at specialised centres—individuals who recognised a problem and sought help. Many people with personality disorders never enter treatment, and those who do in community settings may differ from Kernberg’s population. Generalization should be cautious.
Neurobiological integration is incomplete. While Kernberg’s observations about emotional dysregulation and splitting have been supported by brain imaging research, a full integration of psychodynamic and neuroscience perspectives remains in progress. The biological mechanisms underlying the states Kernberg describes are still being mapped.
How This Research Is Used in the Book
This research is cited in Chapter 2: The Cluster B Conundrum to explain the confusing presentation of patients who show both borderline and narcissistic features:
“The ‘borderline-narcissistic’ combination involves alternating between borderline vulnerability and narcissistic grandiosity. During attachment threat, borderline features dominate: clinging, self-harm, identity confusion. When stabilised, narcissistic defences emerge (devaluation and entitlement). Therapists describe feeling whipsawed between a desperate child and contemptuous adult; interventions appropriate for one state may be harmful in another.”
The citation helps readers understand that personality disorders don’t always fit neat categories—the same person can show dramatically different presentations depending on whether their attachment system feels threatened. This has direct implications for survivors trying to understand their experiences and for clinicians calibrating their interventions.
Historical Context
Kernberg’s 1992 book represents the culmination of his theoretical development regarding aggression in personality pathology. Building on his earlier landmark texts—Borderline Conditions and Pathological Narcissism (1975) and Severe Personality Disorders (1984)—this work placed aggression at the centre of understanding how personality disorders develop and manifest.
Kernberg integrated object relations theory (Klein, Fairbairn) with ego psychology (Hartmann, Jacobson) to create a comprehensive developmental and structural model. His work appeared during a period when psychoanalytic approaches faced challenges from cognitive-behavioural methods, yet demonstrated that depth psychological understanding remained essential for the most severe presentations.
The concepts developed in this book—particularly the borderline-narcissistic combination and the spectrum of narcissistic severity culminating in malignant narcissism—continue to influence both clinical practice and research. Kernberg’s subsequent development of Transference-Focused Psychotherapy created an evidence-based treatment directly derived from these theoretical foundations.
Further Reading
- Kernberg, O.F. (1975). Borderline Conditions and Pathological Narcissism. Jason Aronson.
- Kernberg, O.F. (1984). Severe Personality Disorders: Psychotherapeutic Strategies. Yale University Press.
- Clarkin, J.F., Yeomans, F.E., & Kernberg, O.F. (2006). Psychotherapy for Borderline Personality: Focusing on Object Relations. American Psychiatric Publishing.
- Caligor, E., Kernberg, O.F., & Clarkin, J.F. (2007). Handbook of Dynamic Psychotherapy for Higher Level Personality Pathology. American Psychiatric Publishing.
- Yeomans, F.E., Clarkin, J.F., & Kernberg, O.F. (2015). Transference-Focused Psychotherapy for Borderline Personality Disorder: A Clinical Guide. American Psychiatric Publishing.
About the Author
Otto F. Kernberg, M.D. is one of the most influential psychoanalysts of the modern era and a world authority on severe personality disorders. He is Professor of Psychiatry at Weill Cornell Medical College and Training and Supervising Analyst at the Columbia University Center for Psychoanalytic Training and Research.
Kernberg directed the Personality Disorders Institute at Cornell for over three decades, developing Transference-Focused Psychotherapy (TFP) as an evidence-based treatment for borderline and narcissistic personality disorders. His integration of object relations theory with ego psychology and his systematic approach to treating severe personality pathology have shaped how clinicians worldwide understand and treat these conditions.
He has authored over 20 books and 400 papers, received the American Psychiatric Association's Distinguished Service Award, and continues to teach and supervise well into his 90s. His work on narcissistic personality disorder, particularly the distinction between narcissistic and malignant narcissism, remains foundational to the field.
Historical Context
Published in 1992, this book synthesised Kernberg's decades of clinical work with severe personality disorders. It built upon his earlier foundational texts—Borderline Conditions and Pathological Narcissism (1975) and Severe Personality Disorders (1984)—while focusing specifically on aggression as a central organising principle. The book appeared at a time when psychoanalytic approaches were being challenged by cognitive-behavioural methods, yet it demonstrated that psychodynamic understanding remained essential for the most difficult clinical presentations. It continues to be widely cited in both clinical and research literature on personality disorders.
Frequently Asked Questions
Kernberg explains this through 'splitting'—the inability to integrate positive and negative feelings toward the same person, including oneself. Your abuser likely shifts between idealising/vulnerable states and devaluing/aggressive states because they cannot hold both experiences together. When they feel attachment threat or narcissistic injury, aggression emerges. When they feel secure or need supply, the charming version appears. This isn't acting or manipulation (though it can feel like it)—it's a fundamental failure of psychological integration that makes them genuinely experience the world differently in each state.
Kernberg describes a spectrum. Borderline patients have unstable identity and fear abandonment—they cling desperately to relationships and may harm themselves when threatened with loss. Narcissistic patients have a grandiose identity that defends against a hidden sense of worthlessness—they devalue and discard relationships when those relationships fail to provide admiration. However, many patients show both patterns: borderline vulnerability when attachment feels threatened, narcissistic grandiosity when stabilised. The 'borderline-narcissistic combination' often confuses both clinicians and partners.
No. Kernberg distinguishes between overt aggression (rage, intimidation, violence) and covert aggression (passive-aggression, emotional withdrawal, subtle undermining, exploitation). Covert narcissists may never raise their voice while systematically destroying a partner's self-worth through criticism disguised as concern, chronic disappointment, or emotional unavailability. The aggression is real but hidden—which is why survivors often struggle to explain what was so harmful about relationships that looked 'fine' from outside.
Kernberg believes change is possible but difficult and requires specialised treatment. He developed Transference-Focused Psychotherapy (TFP) specifically for personality disorders, using the therapeutic relationship to help patients integrate split-off parts of themselves. Treatment requires the patient to recognise they have a problem—which narcissistic defences are designed to prevent. Many narcissists never seek treatment, or leave when confronted with their aggression. Those who stay in rigorous treatment for years can achieve meaningful change, but this represents a minority.
Kernberg emphasises that interventions must be carefully calibrated to the patient's current state. Empathic validation appropriate for borderline vulnerability may reinforce narcissistic entitlement. Confrontation appropriate for narcissistic grandiosity may destabilise someone in borderline crisis. Clinicians must track which 'state' is active and adjust accordingly—while also working toward integration so the patient can eventually hold both vulnerability and aggression simultaneously. This requires extensive training and should not be attempted without appropriate supervision.
Kernberg explains this through the concept of splitting. Unable to integrate good and bad qualities in others, narcissists initially see new relationships as 'all good'—idealised sources of admiration and supply. When the inevitable disappointment occurs (no one can maintain perfection), the other person becomes 'all bad'—worthless, contemptible, deserving of aggression. There's no middle ground where someone can be flawed but lovable. The devaluation isn't about you; it's about the narcissist's inability to hold complexity.
Kernberg coined the term 'malignant narcissism' to describe the overlap of narcissistic personality disorder with antisocial features, paranoid traits, and ego-syntonic aggression (aggression the person feels is justified and enjoys). While narcissists generally can feel guilt and have some empathy, malignant narcissists may take pleasure in cruelty and show paranoid vigilance that justifies preemptive aggression. This represents the most severe and dangerous end of the narcissistic spectrum, overlapping with psychopathy.
Critics note that psychoanalytic concepts like 'splitting' are difficult to operationalise and study empirically. The treatment he developed (TFP) requires years and specialised training most clinicians don't have. His framework was developed primarily with patients in long-term intensive therapy—a luxury unavailable to most. Additionally, his understanding is based on clinical inference rather than neuroscience, though subsequent brain imaging research has largely supported his observations about emotional dysregulation in personality disorders.