The relational context of aggress...
The Relational Context of Aggression in Borderline Personality Disorder: Using Adult Attachment Style to Predict Forms of Hostility Kenneth L. Critchfield University of Utah Neuropsychiatric Institute Kenneth N. Levy Pennsylvania State University John F. Clarkin and Otto F. Kernberg Weill Medical College of Cornell University Attachment theory provides a framework for understanding and predicting critical aspects of aggression in the personality disorders. An association between borderline personality disorder (BPD) and insecure forms of adult attachment marked by high relationship anxiety has been repeatedly observed in the empirical literature. Aggression also has been linked to insecure attachment. The present study extends previous work by exploring the degree to which the underlying attachment dimensions of relationship anxiety and avoidance are associated in BPD with the following forms of hostility: (a) direct aggression (verbal or physical) initiated towards others, (b) expectation/perception of aggression from others (including ������reac- tive������ counteraggression when/if provoked), (c) aggression directed towards the self in the form of suicidality or parasuicidality, and (d) affective experience of irritability or anger. The issue was studied in a sample of 92 patients diagnosed with BPD. Results show significant association between more fearful forms of attachment (simultaneous presence of relationship anxiety and avoidance) and the more reactive form of aggression involving expectation of hostility from others. Self- harm was significantly associated only with relational avoidance while This study was supported in part by grants from the Borderline Personality Disorder Research Foundation to principal investigators Otto Kernberg, M.D., and John Clarkin, Ph.D. Correspondence concerning this article should be addressed to: Ken Critchfield, Ph.D., Suite 1648, IRT Clinic, University of Utah Neuropsychiatric Institute, 501 Chipeta Way, Salt Lake City, UT 84108 e-mail: email@example.com JOURNAL OF CLINICAL PSYCHOLOGY,Vol. 64(1), 67--82 (2008) & 2007 Wiley Periodicals, Inc. Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jclp.20434
anger and irritability were associated only with relational anxiety. Implications for understanding relational aspects of BPD aggression in research and clinical work are discussed. & 2007 Wiley Periodicals, Inc., J Clin Psychol 64: 67���82, 2008. Keywords: borderline personality disorder attachment aggression Borderline personality disorder (BPD) is increasingly receiving research attention addressing its etiology, course, and treatment. It is a difficult clinical problem well- known for attendant risks of suicide and self-harm, affective instability, and patterns of idealization and devaluation in relationships. From a diagnostic perspective, BPD poses the problem of having considerable heterogeneity of features. Any combina- tion of five (or more) of nine criteria is required for the diagnosis, producing up to 256 possible variant expressions of the disorder. In addition, BPD shows substantial overlap and comorbidity with all other personality disorders plus Axis I disorders including depression, anxiety, eating disorders, posttraumatic stress disorder, and substance abuse (McGlashan et al., 2000 Nurnberg et al., 1991 Oldham et al., 1992 Zanarini et al., 1998). The usual clinical prototype of BPD involves a picture of impulsive aggression, suicidality and self-harm, extreme dysphoria, abandonment sensitivity, identity disturbance, and unstable, angry affect however, these features do not apply equally to all BPD patients. Heterogeneity suggests that some individuals with BPD are suicidal, some not. Some BPD individuals may be aggressive to others, some not. Among the features thought to be ������core������ to BPD, impulsivity has been identified (Ball, Tennen, Poling, Kranzler, & Rounsaville, 1997 Critchfield, Levy, & Clarkin, 2004 Svrakic et al., 2002 Trull, 1992) as has aggression (Skodol et al., 2002), leading some to hypothesize presence of ������impulsive aggression������ as central to BPD. However, some empirical work has shown that aggression can vary substantially among BPD individuals depending on the precise definition and measure used (Critchfield et al., 2004). Attachment theory provides a useful framework for understanding and predicting critical aspects of personality disorder (Agrawal, Gunderson, Holmes, & Lyons- Ruth, 2004 Crawford et al., 2006 Fonagy, 1999 Levy, 2005). The central observation linking BPD with attachment theory is that characteristic symptoms and problem behaviors are typically in relation to an interpersonal context or are otherwise precipitated by real or imagined events in relationships (Agrawal et al., 2004 Levy, 2005). Attachment theory views relational behavior as reflecting ������internal working models������ or ������internalized representations������ of self and others (Bowlby, 1973). These internalizations are essentially learned cognitive structures and behavioral patterns that develop from perceived early experiences with important attachment figures, usually a child���s parents. Attachment and related internalization of others is thought to be normative and evolutionarily adaptive. From early attachment experiences, a person gains important information about his or her identity and views of self and other as well as capacities for regulating internal experiences and behavioral patterns for maintaining proximity to others. The early learning is potentially modifiable by later learning and experience, but also is thought to shape perception of new events and relationships, thus persisting into adulthood. Experience of aggression from caregivers has been conceptualized as a factor that compromises the ability to develop realistic and balanced views of self and others. Fonagy (1999) summarized, ������in Bowlby���s formulation, aggression���or rather 68 Journal of Clinical Psychology, January 2008 Journal of Clinical Psychology DOI 10.1002/jclp
dysfunctional anger���lies at the root of anxious attachment������ (p. 12). According to theory, insecure attachments may be formed through internalization of abuse and neglect in the form of a relational template with potential to guide expectancies for similar abuse and/or rejection in future relationships (Critchfield & Benjamin, in press Florsheim, Henry, & Benjamin, 2006 Fonagy, 1999 Levy, 2005). However, not all patients diagnosed with BPD have obvious abuse, neglect, or rejection in their histories, leading some to posit that high levels of constitutionally based aggression or negative affect also may interfere with early learning in such a way as to prevent an integrated view of self and others, ultimately producing a similar template for relational behavior containing strong elements of hostility and aggression (Gurvits, Koenigsberg, & Siever, 2000 Kernberg, 1984 Levy et al., 2006). Consistent with all these theories, an association between BPD (often characterized by early trauma as well as high levels of negative affect) and insecure forms of adult attachment marked by high relationship anxiety have been repeatedly observed in the empirical literature (Agrawal et al., 2004 Levy, Meehan, Weber, Reynoso, & Clarkin, 2005).1 Bartholomew and Horowitz (1991) first characterized the fearful style of adult attachment as having a negative view of the self in combination with negative views of close others. Florsheim et al. (1996) used an interpersonal framework to further refine the description, defining the fearful style in part as having expectations that close others will attack, reject, or blame. Consistent with these perspectives, Dutton (2002) noted that a frequent motivation for relational aggression is belief that the partner will abandon the abuser. Empirical work conducted by Dutton and colleagues (Dutton, 2002 Dutton, Saunders, Starzomski, & Bartholomew, 1994) has shown adult aggression against romantic partners to be associated with insecure attachment styles, particularly the fearful type���so much so that the term ������angry attachment������ has been proposed as an alternate descriptor. The same researchers also reported high levels of BPD in their partner-abusing samples. Recent work has found that aggression is a useful concept for understanding the personality disorders when added to considerations of attachment style (Brennan & Shaver, 1998 Crawford et al., 2006 Levy, 2005). Crawford and colleagues (2006) found that Cluster B disorders are associated with high levels of relationship anxiety, avoidance, and aggression, with aggression distinguishing Cluster B from other relationally anxious personality disorders such as avoidant and dependent. They concluded that assessment of aggression may be necessary to differentiate personality disorders over and above information available from attachment style alone. They also called for more fine-grained work to understand the relation between aggression and attachment at the level of individual personality disorders. The present study extends the work of Crawford et al. (2006) by exploring the degree to which individuals diagnosed with BPD exhibit particular attachment styles, and the degree to which the underlying attachment dimensions of relationship anxiety and avoidance are themselves associated with forms of aggression. We expect to replicate previous findings of high levels of insecure attachment in BPD, especially attachments marked by relational anxiety (i.e., preoccupied and fearful). Rather than assume that all forms of aggression are equivalent, we examine the construct separately for: (a) direct aggression (verbal or physical) initiated toward others, (b) expectation/perception of aggression from others (which includes ������reactive������ counteraggression when/if provoked), 1 See Levy (2005) for a thorough review of theory and research on attachment in BPD. 69 Attachment and Aggression in BPD Journal of Clinical Psychology DOI 10.1002/jclp